Trajectories of rehabilitation exercise adherence after hip fracture surgery in older adults: a prospective longitudinal study

Recovery after hip fracture surgery in older adults depends on sustained rehabilitation exercise, yet adherence after discharge is heterogeneous and may change over time. Previous studies have focused on single time points or average changes, which may overlook clinically meaningful heterogeneity. This study aimed to identify trajectories of rehabilitation exercise adherence during the first 6 months after hip fracture surgery and to examine factors associated with trajectory membership. This prospective longitudinal study enrolled 199 older adults undergoing hip fracture surgery at a secondary-level Class A general hospital in Hangzhou, China, June–December 2025. Baseline demographic, clinical, and psychosocial data were collected during hospitalization. Rehabilitation exercise adherence was assessed at 1, 3, and 6 months postoperatively using the Orthopaedic Patient Functional Exercise Compliance Scale. Latent growth mixture modelling identified adherence trajectories. Univariate analyses and multinomial logistic regression were used to examine factors associated with trajectory membership, with binary logistic regression performed as a sensitivity analysis. Follow-up assessments included 196 participants at 1 month, 194 at 3 months, and 191 at 6 months. Mean adherence scores increased steadily, from 30.00 ± 16.49 at 1 month to 38.39 ± 14.95 at 3 months and 48.58 ± 18.85 at 6 months. A 3-class solution was selected as optimal and clinically interpretable, identifying a low-adherence trajectory (11/199, 5.53%), a moderate-adherence trajectory (155/199, 77.89%), and a high-adherence trajectory (33/199, 16.58%). Trajectory groups differed significantly in education, marital status, household income, social support, self-efficacy, and rehabilitation plan type. Multinomial logistic regression showed that higher education, household income, and self-efficacy were associated with lower odds of low or moderate adherence, whereas a high-intensity rehabilitation plan was associated with higher odds of lower adherence. A binary sensitivity analysis comparing non-high versus high adherence yielded directionally consistent associations. Older adults in this single-centre cohort showed heterogeneous patterns of rehabilitation exercise adherence over 6 postoperative months. Education, household income, exercise self-efficacy, and rehabilitation plan type were associated with adherence patterns, although regression findings were exploratory. These findings require confirmation in larger multicentre studies and may inform individualized rehabilitation planning.